| 2 Oct 2025 |
Puna | * as-is, qemu seems unable to launch a regular kernel with the powernv platform selected. i only managed to get something resembling a booting system out of that platform by passing petitboot files into -kernel and -initrd, and adding a CD device with an installer ISO. since we also use those two options to start VMs for tests though, that doesn't really work - at least not without introducing some nix code to make a petitboot-recognised dis(c/k) image to pass via a device… | 17:42:56 |
Puna | just curious if you have any more insight into / ideas about that, was hoping to do installer tests for IEEE-1275 grub2 before committing to writing anything onto a real disk :) | 17:45:48 |
Puna | * the pseries platform can definitely use -kernel and -initrd to boot a regular linux kernel, but on a big-endian POWER system any kernel I try to boot inside of the VM just runs into invalid memory access almost immediately. so that also doesn't really work, at least not for natively running the tests | 17:47:47 |
Puna | * the pseries platform can definitely use -kernel and -initrd to boot a regular linux kernel, but on a big-endian POWER system any kernel I try to boot inside of the VM just runs into invalid memory access almost immediately. so that also doesn't really work, at least not for natively running the tests on the platform that I have access to | 17:48:07 |
lzcunt | Heyy, it looks like bash introduced a sanity check for strchrnul that fails when cross compiling (just like getenv) and bash tries to replace the implementation | 20:16:54 |
lzcunt | but for static builds with mlibc it fails because mlibc doesn't have them as weak symbols (which is similar behaviour to BSDs) | 20:17:53 |
lzcunt | someone working with cross compilation to BSDs should check if they're affected | 20:18:23 |
lzcunt | if they're affected, bash_cv_func_strchrnul_works=yes when cross compiling to BSDs just like we currently do for getenv_works should be enough to fix it | 20:19:51 |
Alyssa Ross | In reply to @sananatheskenana:matrix.org someone working with cross compilation to BSDs should check if they're affected BSD cross works well enough that it should be quite easy to check yourself if you're interested (and want to also fix it for mlibc) | 20:22:44 |
lzcunt | I have it fixed for mlibc locally | 20:23:01 |
lzcunt | I'll check then I'll get them fixed in the same commit then? | 20:23:14 |
lzcunt | Sorry for the slow progress on the mlibc PR btw, I was moving and I got into uni | 20:24:00 |
lzcunt | Now I'm doing PRs in 0-credit classes | 20:24:21 |
lzcunt | Checked, strchrnul appears to be a weak symbol on freebsd | 20:57:38 |
lzcunt | And static bash cross compiles | 20:57:54 |
| 3 Oct 2025 |
| Venticx joined the room. | 10:09:37 |
| 4 Oct 2025 |
| aleksana 🏳️⚧️ (force me to bed after 18:00 UTC) changed their profile picture. | 08:41:25 |
| 5 Oct 2025 |
Puna |  Download image.png | 15:10:19 |
Puna | booting fails, and display never gets any output, but baby steps 🙂 | 15:11:08 |
Puna | fails in the emulation cus the installer's kernel can't find /dev/root, but it works on my hardware. maybe something works different on pseries machines there, hmm… | 22:20:45 |
Puna |  Download IMG_20251006_001633.jpg | 22:20:57 |
Puna | :3 | 22:21:09 |
Puna | * fails in the emulation cus the installer's kernel can't find /dev/root, but it works on my hardware. maybe something works different on pseries machines there, or qemu just behaves weird somewhere. hmm… | 22:21:46 |
| 6 Oct 2025 |
pandapip1 | Has there been any effort to port openwrt/pmos kernels & stuff for nixos | 03:12:12 |
K900 | Generally we try to minimize the delta to mainline | 04:44:46 |
K900 | e.g. I have a bpi-r4 running with some openwrt-adjacent patches | 04:44:59 |
K900 | But it's like 50 commits on top of mainline and some are backports from net-nexr | 04:45:12 |
K900 | * | 04:45:17 |
Alyssa Ross | In reply to @k900:0upti.me Generally we try to minimize the delta to mainline (so does pmos) | 06:05:49 |
| 11 Oct 2025 |
| Anders joined the room. | 12:34:28 |